RESEARCH ARTICLE Open Access

Factors associated with lung cytology asobtained by non-endoscopic broncho-alveolar lavage in group-housed calvesKatharina van Leenen1* , Laura Van Driessche1, Lieze De Cremer1, Linde Gille1, Christien Masmeijer1, Filip Boyen2,Piet Deprez1 and Bart Pardon1

Abstract

Background: Respiratory infections are the main indication for antimicrobial use in calves. As in humans andhorses, studying inflammation of the deep airways by lung cytology raises the possibility of preventing respiratorydisease and targeting its treatment in the future. Whether lung cytology findings coincide with clinical signs andlung ultrasonographic findings is currently unknown. Therefore, the objective of the present study was todetermine the association of lung cytology with clinical signs, lung consolidation and broncho-alveolar lavage fluid(BALf) characteristics (including bacteriology).A total of 352 indoor group-housed calves aged between 1 and 6months from 62 conveniently selectedcommercial herds were included in this cross-sectional study. Clinical examination, thoracic ultrasound andbacteriology and cytology on non-endoscopic broncho-alveolar lavage (nBAL) samples were performed.

Results: Pneumonia, defined as presence of ultrasonographic lung consolidations ≥1 cm in depth, affected 42.4%of the calves. Mean BALf neutrophil percentage was 36.6% (SD 23.8; R 0–97.4) and only a positive induced trachealcough reflex (P = 0.04), standing posture (P = 0.03) increased breathing rate (P = 0.02) and isolation of Pasteurellamultocida (P = 0.005), were associated with increased neutrophil percentage. No significant associations betweenlung ultrasonographic findings and cytology results were present, except for presence of basophils in BALf andconsolidation of > 3 cm in depth (OR = 2.6; CI = 1.2–5.6; P = 0.01). Abnormal lung sounds were associated withdetection of eosinophils in BALf (OR = 2.8; CI = 1.0–8.1; P = 0.05). Total nucleated cell count (TNCC) (P < 0.001) waspositively and macrophage percentage (P = 0.02) negatively associated with volume of lavage fluid recovered.Macroscopic blood staining of BALf increased TNCC (P = 0.002) and lymphocyte percentage (P = 0.001).

Conclusions: Only a limited number of clinical signs and ultrasonographic findings were associated with nBALcytology. BALf cytology offers additional and distinct information in calves aiding in detection and prevention ofrespiratory conditions. In this population, selected from herds not reporting any recent respiratory illness, a highnumber of calves had ultrasonographic lung consolidation and high neutrophil percentage in BALf, suggesting thatsubclinical disease presentations frequently occur.

Keywords: Neutrophilia, Respiratory disease, Total nucleated cell count, Eosinophils, Thoracic ultrasound, Pasteurellamultocida

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

* Correspondence: Katharina.vanLeenen@UGent.be1Department of Large Animal Internal Medicine, Faculty of VeterinaryMedicine, Ghent University, Salisburylaan 133, 9820 Merelbeke, BelgiumFull list of author information is available at the end of the article

van Leenen et al. BMC Veterinary Research (2019) 15:167 https://doi.org/10.1186/s12917-019-1921-x

BackgroundRespiratory tract infections have severe economic conse-quences for the cattle industry worldwide [1–3]. In re-cent years, their importance has even increased becausethey are the main reason for antimicrobial use in calvesand feedlot animals [4, 5]. Reduction of antimicrobialuse in food producing animals is a top priority of theEuropean Union [6]. Respiratory tract infections induceinflammation of the upper, and in many cases also thelower respiratory tract [7–9]. However, besides infectionsairway inflammation can be caused by non-infectiousfactors, such as air pollutants, related to bad stable envi-ronments [10–12]. Airway inflammation of various ori-gin can reduce pulmonary function, induce mucusaccumulation and reduce ciliary activity, potentiallyresulting in secondary (bacterial) pneumonia, as demon-strated in humans and horses [10, 13]. In cattle, studyingairway inflammation is largely unexplored territory. Not-withstanding, knowledge on airway inflammation hasthe potential to help the industry towards a more pre-ventive approach and to better targeted antimicrobialtreatment. The reference method to study inflammationof the deep airways, is cytology on broncho-alveolarlavage (BAL) samples [14–17]. This can be done bybronchoscopic methods or by the use of non-endoscopictechniques, using a BAL-catheter [18, 19]. Both BALtechnique and flushing volume greatly influence cyto-logical results [17, 20]. For calves, available publicationson lung cytology only deal with bronchoscopic BALtechniques, in a limited number of animals [8, 21].However, in several Western European countries, veteri-narians increasingly use a non-endoscopic BAL method(nBAL). Sterilisable custom-made or commercially avail-able BAL catheters are used, allowing sampling of mul-tiple animals in a limited time frame at low cost [22].This blind technique does not systematically sample thediaphragmatic lung lobes, but a random lung lobe [23].Today, practitioners use this sampling technique toidentify pathogens involved in outbreaks of respiratorydisease, despite concern on possible contamination ofnBAL samples by nasal passage [24]. Nevertheless, incontrast to bronchoscopic procedures, nBAL has foundits way to the field, opening possibilities to include lungcytology as an additional parameter for the managementof respiratory diseases in calves. For welfare reasons,veterinarians prefer a small flushing volume (30 mL) fornBAL [22]. This type of mini-BAL has recently alsogained access in human medicine for both lung cytologyas bacteriology for example in the management of venti-lator associated pneumonia or in immunocompromisedpatients [25–27]. To what extent lung cytology coincideswith clinical signs or the more recently introducedon-farm lung ultrasonographic findings [28, 29], iscurrently unexplored. Whether nBAL cytology and

bacteriology results are associated is also unknown, inanimals as in humans.The objective of the present study was to determine

whether clinical signs, lung ultrasound, and broncho-al-veolar lavage fluid (BALf) characteristics (includingbacteriology) are associated with lung cytology of nBALsamples in a population of indoor group-housed calves.

MethodsSample size calculation, study design, and animalsThe sample size required to detect a 10% difference inneutrophil percentage between calves with and withouta given risk factor, using 5% as a reference as seen inhealthy horses [16], was 139 animals in each test group(with 95% confidence and 80% power). Sample size wasfurther increased to 352 animals, to assure inclusion ofthe most common bacterial pathogens in the sample andthe possibility to explore factors with multiple categor-ies. A cross-sectional field study was performed on 62conveniently selected commercial herds (23 dairy, 23beef, 14 mixed and 2 veal) between January and April2017. Herds were conveniently selected, with the help ofdifferent local veterinary practices, on willingness to co-operate and covered mainly the provinces West and EastFlanders (Belgium). The only inclusion criteria was ab-sence of an epidemic episode of respiratory disease inthe last 2 months to avoid massive neutrophilia in themajority of the samples. An epidemic episode of respira-tory disease has been defined as 20% new cases of re-spiratory disease in the same stable or age category in a24-h period. Animal selection criteria were indoorgroup-housing, absence of oral or systemic antimicrobialtreatment in the past 2 weeks and age between 1 and 6months. The objective was to sample 8–10 calves perfarm, housed in the same pen or 2 adjacent pens. If lessanimals were present all calves were sampled, if more than10 calves were present in one pen animals were selectedrandomly. Veal calves were group-housed on a slattedfloor and fed milk replacer, concentrates and roughage ac-cording to European legislation (EC2008–119). Beef anddairy calves were both group-housed on straw beddingand fed milk replacer, concentrates and roughage withsubstantial variation between farms, all calves had ad libi-tum access to water. All sampling techniques and thestudy protocol were revised by the Ethical Committee ofthe Faculty of Veterinary Medicine, Ghent University andpermitted under experimental licence number EC2016–89. After the study all calves remained on the farm, noanimals were sacrificed for this study.

Clinical examination and thoracic ultrasonographyFor each calf data on the following clinical signs was col-lected: mentation (alert versus depressed; depressed =decreased activity, eye closure, reduced awareness of the

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 2 of 12

environment, lowering of the head), posture (standing;sternal recumbency; lateral recumbency), head tilt(present; absent), position of the ears (normal; unilat-erally drooped or bilaterally drooped), nasal discharge(absent; unilateral; bilateral), ocular discharge (absent;unilateral; bilateral), type of ocular or nasal discharge(serous; seromucous; mucopurulent; purulent), spontan-eous cough (present; absent), breathing frequency (inbreaths per minute), heart rate (in beats per minute),rectal temperature (°C), induced laryngeal cough reflex(positive; negative), induced tracheal cough reflex (posi-tive; negative), faecal consistency (normal; pasty; waterydiarrhea) and lung auscultation (normal versus abnor-mal; abnormal = increased respiratory sounds, pleuralfriction sound or presence of wheezes or crackles). Apositive induced laryngeal or tracheal cough reflex wasdefined as a single induced cough following manualcompression of the larynx or trachea, respectively.Thoracic ultrasound was performed with a linear

probe with a frequency of 7.5-MHz (Tringa Linear Vet®,Esaote, the Netherlands), set at 8 cm of depth, usingisopropyl alcohol (70%) as a transducing agent as previ-ously described [30]. Presence of lung lesions was docu-mented according to location (dorsal;ventral / left;right)and size of the lesion. Size was categorized using anadapted 6-point scale ultrasonographic lesion score(ULS). The definitions used for the ultrasonographiclesion score were as proposed by Ollivett et al., [21] withslight modifications: normal (ULS 0): only normal rever-beration artefacts; comet tails (ULS 1): < 8 comet-tailartefacts in one field without presence of hypoechoicconsolidations; interstitial syndrome (ULS 2): diffuse (>8) comet-tail artefacts in one field without presence ofhypoechoic consolidations; small consolidation (ULS 3):hypoechoic consolidation < 1 cm in depth; moderateconsolidation (ULS 4): hypoechoic consolidation 1–3 cmin depth; severe consolidation (ULS 5): hypoechoicconsolidation > 3 cm in depth. Consolidation depth wasmeasured in a dorso-ventral plane using the grid on thescreen of the ultrasound. Pleural effusion, characterizedas a line of hypoechoic fluid between the lung and thepleural interface, was described as being absent orpresent.

Broncho-alveolar lavage technique and BALf analysisBALf was collected by a non-endoscopic technique usinga reusable custom made polytetrafluorethylene catheterof 1.5 m length with a 12G- catheter stylet and an innerand outer diameter of 2 mm and 4mm, respectively(VWR, Leuven, Belgium). The procedure was performedin standing, unsedated animals by the same veterinarian,using a new sterilized catheter for each calf as describedpreviously [23]. Briefly, after the nostril was cleaned with70% isopropyl alcohol the catheter was inserted into the

ventral meatus of the nose and advanced to the bronchiuntil the wedge position was reached. One aliquot iso-tonic sterile saline was instilled using a volume of 0.6ml/kg body weight. After instillation the fluid was imme-diately aspirated and if no fluid was recovered, another20 mL of saline was instilled. The volume of instilledand recovered BALf and macroscopically visible bloodstaining of the sample were recorded for each animal. Ifthe esophagus was accidently entered during the attemptto access the trachea a new sterilized catheter was usedand the sampling process repeated from the start toavoid contaminated samples. Samples were transportedin plastic tubes on ice and processed within 12 h aftersampling.Total nucleated cell count (TNCC) of the recovered

lavage fluid was determined manually using a haemocyt-ometer. The sample was vortexed and 1 μL of BALf wasdiluted with 10 μL Türk’s solution (Merck KGaA, Darm-stadt, Germany) and counted manually using a Bürkercounting chamber (Marienfeld GmbH & Co. KG,Lauda-Königshofen, Germany). Diff-Quick (MerckKGaA, Darmstadt, Germany) stained cytocentrifuge(Shandon Scientific, London, UK) preparations of BALf(1200 rpm for 10 min) were made and a total of 400nucleated cells was counted at × 100 magnification tocalculate the differential cell count [18]. All specimenswere counted by the same observer (trainedveterinarian).Bacterial culture was performed using Columbia blood

agar enriched with 5% sheep blood (Oxoid™, Hampshire,UK) for isolation of Pasteurellaceae and pleuropneumonia-like organism agar (PPLO) (Difco™, BectonDickinson and Company, Franklin Lakes, NJ, USA), forisolation of Mycoplasmataceae. Incubation was done at35 °C and 5% CO2, overnight and for 5 days, for Pasteur-ellaceae and Mycoplasmataceae, respectively. Myco-plasma bovis was identified by inoculating colonies on aPPLO agar enriched with polysorbate 80 (Difco™, BectonDickinson and Company, Franklin Lakes, NJ, USA),followed by microscopic identification of growing col-onies by their typical morphology [31]. Species confirm-ation of Pasteurellaceae was performed usingMatrix-Assisted Laser Desorption Ionization-Time ofFlight Mass Spectrometry (MALDI-TOF MS) (BrükerDaltonik GmbH, Bremen, Germany). Bacterial culturesfor Pasteurellaceae were interpreted as negative, polymi-crobial and dominant or pure cultures of Pasteurellamultocida, Mannheimia haemolytica and Histophilussomni as described previously [22]. M. bovis cultureswere interpreted as positive or negative.

Statistical analysisAll data were entered in a spreadsheet (Excel, MicrosoftInc. Washington, USA) and transferred to SAS 9.4 (SAS

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 3 of 12

Institute Inc., Cary, N.C, USA) for statistical analysis. Todetermine the relationship between TNCC and thedifferent cell populations (in %) scatter plots were madeand linear regression applied (PROC REG). Outcomevariables were checked for a normal distribution and log+ 1 transformed when required. Samples were consid-ered positive for eosinophils if > 1% eosinophils werepresent in BALf and were considered positive for baso-phils if any of these cells were present in the 400 cellscounted [32, 33].Five multivariable linear regression models (PROC

MIXED) were made with TNCC, neutrophil percentage,macrophage percentage, lymphocyte percentage, andpercentage of epithelial cells as outcome variables.Predictors were clinical signs, BALf characteristics andultrasonographic parameters. Thirteen clinical signswere tested: mentation (alert; depressed), posture (stand-ing; sternal recumbency), head tilt (present; absent), earposition (normal; unilateral droopy ears; bilateral droopyears), nasal discharge (absent; unilateral; bilateral), oculardischarge (absent; unilateral; bilateral), spontaneouscough (present; absent), breathing frequency, heart rate,temperature, induced tracheal cough reflex (negative;positive), induced laryngeal cough reflex negative; posi-tive) and lung auscultation (normal; abnormal). Six BALfcharacteristics were tested: volume of BALf recovered(in mL), blood staining (absent; present), P. multocidaisolation (negative; positive), M. haemolytica isolation(negative; positive), H. somni isolation (negative;positive), and M. bovis isolation (negative; positive). Fourdifferent binary outcomes were created based on ultra-sonographic findings: ultrasonographically normal lung(reverberation artefacts and < 8 comet-tail artefacts inone image), diffuse comet-tail artefacts (> 8 in oneimage), lung consolidation (if a consolidation waspresent the depth in centimetres was noted), pleural ef-fusion (absent; present). In each model herd was addedas a random effect to account for clustering of calveswithin a herd. In a first step the association of the differ-ent predictors with the outcome variable was tested uni-variably. Continuous parameters (temperature, breathingrate, heart rate and BALf volume) were tested bothcontinuously and categorically based on quartiles and re-ceiver operating characteristics curve with the Youden’sindex to determine optimal cut-off to create a binaryvariable. All parameters with P < 0.20 were withheld forthe next step. Pearsons and Spearman correlation weredetermined, and of predictors correlated over 0.6, onlythe most significant one was withheld for the multivari-able model. The multivariable model was built stepwisebackwards, gradually excluding not significant variables.Significance was set at P < 0.05 and 0.05 < P < 0.10 wasconsidered a trend. Pairwise comparisons between differ-ent categories of significant effects were made using

Bonferonni corrections. Biologically plausible interac-tions between significant main effects were tested.Model fit was assessed by visual inspection of residualplots and normality testing of residuals.For BALf eosinophils and basophils a multivariable

logistic regression was used. Samples containing > 1%eosinophils were considered positive (increased), as inhumans and horses [32, 33]. For basophils, a sample wasconsidered positive if any of these cells were seen in the400 cells counted [33]. A generalized linear mixed model(PROC GLIMMIX) was used with binomial distributionand logit link function with Wald’s statistics for type 3 con-trasts. Herd was added as a random factor to account forclustering. First, the same predictors as mentioned abovewere tested in univariable analysis. The same selection andsignificance criteria as mentioned above were used for thefurther model building procedure. Model fit was evaluatedusing the Hosmer-Lemeshow goodness-of-fit test for logis-tic models. All possible and biologically relevant interac-tions between significant main effects were tested.

ResultsAnimals and herdsOn 62 farms a total of 352 animals were sampled,consisting of 50.1% (178/352) Holstein-Friesian calves,44.3% (156/352) Belgian Blue calves and 5.1% (18/352)mixed breed calves. Of the Holstein-Friesian calves 5%(9/178) originated from a veal calf facility, the remaining95% (169/178) from a dairy farm. On 87.1% (54/62) ofthe farms less than eight calves were present in one penor in adjacent pens. Therefore all eligible calves weresampled on these farms. None of the farms had ten ormore calves in one pen or adjacent pens. Of the calves,4.3% (15/352) were aged 4 weeks or less, 35.8% (126/352) were aged between 4 and 8 weeks and 59.9% (211/352) of the calves were more than 8 weeks old.

Clinical signs and ultrasonographic lesionsMean rectal temperature was 39.0 °C (standard deviation[SD] = 0.5; range [R] = 36.6–41.2) and mean breathingfrequency 35 breaths/minute (SD, 24; R, 14–116). In51.7% (182/352) of the animals at least one of the re-corded clinical signs was present. Uni- or bilateral nasaldischarge was the most frequently observed sign, presentin 22.2% (78/352) of the calves, followed by spontaneouscough in 21.0% (74/352), uni- or bilateral drooped earsin 5.7% (20/352) and depression in 4.2% (15/352). Apositive tracheal reflex could be induced in 14.2% (50/352) of the animals whereas the laryngeal reflex waspositive in only 2.6% (9/352). Thoracic ultrasound dem-onstrated lung consolidation with a depth of ≥1 cm(ULS 4–5) in 42.4% (149/352) and a consolidation > 3cm (ULS 5) in 23.9% (84/352) of the calves. Pleuraleffusion was seen in 4.5% (16/352) of the calves. The

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 4 of 12

anatomical distribution of the ultrasonographic lesions ispresented in Fig. 1. No significant age, breed or herd ef-fects on the presence of lung consolidations (≥1 cm indepth) could be demonstrated. For lung consolidations> 3 cm in depth, significant breed (dairy > beef breeds(odds ratio (OR) = 5.6 (95% confidence interval (CI) =2.0–16.2; P < 0.001) and herd effects (P < 0.001) couldbe demonstrated, but not age effects (P = 0.14).

BALf characteristics, cell counts and bacteriologyThe average percentage of instilled fluid was 38.8 ml(SD, 5.2 ml; R, 25–80ml) and in 1.7% (6/352) of thecalves an additional 20 ml of saline needed to be instilleddue to insufficient recovery. Mean percentage of instilledfluid recovered was 33.5% (SD, 9.9; R, 12.3–73.8) and11.1% (39/352) of the samples were macroscopicallyblood stained. Blood stained samples were significantlymore frequent in calves aged > 8 weeks (14.7%; [31/211])compared to calves aged between 4 and 8 weeks (5.6%;[7/126]) (P = 0.04).Mean TNCC was 1.9 × 109 cells/L (SD, 1.7; R, 0–13.7).

The ULS score was not associated with TNCC. MeanBALf differential cell percentage was 42.8% macrophages(SD, 18.9; R, 2.4–92.3), 36.6% neutrophils (SD, 23.8; R,0–97.4), 5.4% lymphocytes (SD, 5.2; R, 0–45.8), 0.3% eo-sinophils (SD, 0.8; R, 0–9.1), 0% basophils (SD, 0.1; R,0–0.7) and 14.9% epithelial cells (SD, 13.0; R, 0–95.9), as

shown in Additional file 1: Table S1. Of the calves 5.7%(20/352) showed eosinophilia (> 1% eosinophils) and10.5% (37/352) had basophils in their BAL sample.Calves with eosinophils and basophils in their BALf werepresent on 25.8% (16/62) and 35.5% (22/62) of thefarms, respectively.A positive correlation was found between neutrophil

percentage and TNCC (r = 0.20) whereas macrophagepercentage was negatively correlated with TNCC (r =0.13). Between the percentage of epithelial cells and per-centage of neutrophils in BALf a negative correlationwas noted (r = 0.28) (Fig. 2). No other significant correla-tions between the different cell types were present. Nosignificant differences were found between breed, age,clinical signs and BALf cellular characteristics, as shownin Additional file 2: Table S2. No significant differenceswere found between BALf differential cell count of lungswithout ultrasonographic consolidations versus lungswhich did show these ultrasonographic lesions or pleuraleffusion.A pure or a dominant culture for Pasteurellacea were

present in 16.8% (59/352) and 33.5% (118/352) of theBALf samples, respectively. In 40.6% (143/352) of thesamples a polymicrobial result was obtained and 9.0%(32/352) of the BALf samples were bacteriologicallynegative. P. multocida was isolated from 31.2% (110/352) of the samples, M. haemolytica from 14.2% (50/

Fig. 1 Anatomical distribution of lung lesions grouped by ultrasound lesion score (ULS) from 352 indoor group-housed calves. ULS 0 = onlynormal reverberation artefacts; ULS 1 = < 8 comet-tail artefacts in one image present; ULS 2 = diffuse (> 8) comet-tail artefacts without presence ofhypoechogenic consolidations; ULS 3 = hypoechogenic consolidation < 1 cm in depth; ULS 4 = hypoechogenic consolidation 1–3 cm in depth;ULS 5 = hypoechogenic consolidation > 3 cm in depth

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 5 of 12

352), H. somni from 2.3% (8/352) and 3.1% (11/352) ofthe samples were positive for M. bovis. Mixed infectionswere present in 5.1% (18/352) of the samples and con-sisted of M. bovis and P. multocida (1.4% [5/352]), P.multocida and M. haemolytica (1.4% [5/352]), M. bovisand M. haemolytica (1.1% [4/352]), M. haemolytica andH. somni (0.6% [2/352]) and P. multocida and H. somni(0.6% [2/352]) combinations. Pathogen isolation rateswere not linked with clinical signs or ultrasonographicfindings.

Factors associated with TNCC and differential cell countsLog transformation was needed for TNCC, neutrophilpercentage, lymphocyte percentage and epithelial cellpercentage. Macrophage percentage was normally dis-tributed. For eosinophil and basophil percentage notransformation to a normal distribution was possible,hence it was opted to analyse them as binary outcomes.

When performing the final mixed model analysis 17calves were excluded for the analysis of BALf neutro-phils, lymphocytes and epithelial cells and 20 calves wereexcluded for the analysis of BALf TNCC and macro-phages, due to incomplete information on some of thepredictors studied, resulting in 335 and 332 calves forfinal analyses, respectively.The final multivariable models for TNCC (Table 1),

neutrophil percentage (Table 2), macrophage percentage(Table 3), lymphocyte percentage (Table 4), epithelial cellpercentage (Table 5), presence of > 1% eosinophils andpresence of basophils (Table 6) are available in the re-spective tables. Older calves (> 8 weeks) had a signifi-cantly higher lymphocyte percentage, but no age effectson other cell types were noticed (Table 4). For BALfcharacteristics, the higher the recovered BALf volume,the higher TNCC and the lower the macrophage per-centage were (Tables 1 and 3). Macroscopic blood

Fig. 2 Linear associations between different broncho-alveolar lavage fluid cellular components from 352 indoor group-housed calves. Significantassociations were present for total nucleated cell count and neutrophil percentage (a), total nucleated cell count and macrophage percentage (b)and epithelial cell percentage and neutrophil percentage (c). No significant associations were found between other cell types

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 6 of 12

staining of BALf, significantly increased TNCC andlymphocyte percentage and was associated with thepresence of basophils (Tables 1, 4 and 6). Of all clinicalsymptoms recorded only a positive induced trachealcough reflex, standing position of the calf and increasedbreathing rate remained associated with cellular changesin BALf in the multivariable models. A positive inducedtracheal reflex was associated with an increased TNCCand neutrophil percentage and decreased epithelial cellpercentage (Tables 1, 2 and 5). An increased breathingfrequency was associated with a decreased percentage ofepithelial cells and an increased neutrophil percentage(Tables 2 and 5). Clinical signs frequently used to iden-tify calves suffering from respiratory disease such asnasal and ocular discharge, fever and spontaneous coughwere not associated with any of the studied outcomes.Of the upper quartile of calves with the highest BALfneutrophil percentage 21.6% (19/88) had an increasedbreathing frequency (> 44 bpm) and 19.3% (17/88) apositive induced tracheal reflex.For pathogen isolation (Fig. 3), only a positive P. mul-

tocida culture was associated with increased BALf neu-trophil percentage and a reduction in macrophage

percentage (Tables 2 and 3). H. somni isolation was asso-ciated with detection of basophils (Table 6). When test-ing the ULS as a categorical factor, no associations withany of the outcomes studied was found. When using dif-ferent cut-offs to categorize the depth of ultrasono-graphic lung consolidation (> 1 cm; > 3 cm; > 6 cm), onlyconsolidations with a depth of > 1 cm were univariablyassociated (P = 0.04) with an increased BALf neutrophilpercentage. Only for the presence of basophils a signifi-cant association with lung consolidation > 3 cm (P =0.001) remained in the multivariable model (Table 6).

DiscussionIn this study the association of clinical signs and lungultrasonography with BALf cytology was explored, togain insights in the potential added value of cytology forrespiratory disease diagnostics and prevention in futurework.To include as much variation in the cytological vari-

ables as possible these associations were studied in apopulation of animals housed in herds with onlyendemic respiratory problems. Herds with endemicrespiratory problems were defined as herds with no

Fig. 3 Total nucleated cell count and differential cell counts in broncho-alveolar lavage fluid from 352 indoor group-housed calves by isolatedpathogen. a = total nucleated cell count, b = neutrophils, c =macrophages. Negative = no major pathogens could be isolated from the broncho-alveolar fluid; P.m = Pasteurella multocida isolation; M.h =Mannheimia haemolytica isolation; H.s = Histophilus somni isolation

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 7 of 12

epidemic respiratory disease presentation in the last 2months. Hence, the finding that in 42.4% of the calveslung consolidations (≥1 cm in depth) were present, des-pite that half of the calves studied did not show any ab-normal clinical signs, was unexpected. This prevalence ishigher than previously reported for 60-day-old Holsteinheifers (20.2% (any consolidation) [34] and three-month-old Jersey heifers 27.6% (consolidation ≥1 cm)[1]. Notably, severe ultrasonographic lesions were morefrequently present in Holstein-Friesian calves, whereasthe Belgian blue breed has been known to be more sus-ceptible for respiratory disease [35]. Possible explana-tions could be that dairy calves expressed clinical signsof pneumonia to a lesser extent, whereas beef farmerscould have observed their calves closer and treated themmore intense, given the greater value of the animals.Furthermore, the possibility of less awareness of (sub)-clinical pneumonia in dairy calf farming in this regioncannot be excluded. Besides breed, differences in man-agement, housing and prophylactic regimens could haveaffected the prevalence of lung consolidation, neverthe-less, this was outside the scope of this article.This study showed that in this population of animals,

housed in herds with only endemic respiratory problems,no association of lung ultrasonographic findings andlung cytology was found. This is in contrast to previouswork on experimental animals using an endoscopic BALtechnique [21]. Most likely these discordant findings are

related to sampling of a random lung lobe with thenBAL technique, which could have resulted in samplingof a healthy lobe instead of a consolidated one in someanimals. With the endoscopic lavage technique theconsolidated lung lobe can be selectively sampled. Anequally important finding was that not all of the com-monly monitored respiratory signs were associated withinflammation (higher neutrophil percentage) of thelower airways. Only the induced tracheal cough reflex,posture of the calf and increased breathing rate werelinked to neutrophil percentage. These signs might bemore specific for lower respiratory tract inflammation,and potentially infection, which could stress their poten-tial benefit for identification of animals requiring treat-ment, in contrast to signs like nasal discharge or fever.A tracheal cough reflex is triggered by stimulation of theirritated and inflamed mucosa of the lower airways,characterized by an influx of neutrophils. Calves thatshowed increased respiratory rates might have beensuffering from hypoxia resulting from a more advancedpneumonia or acute interstitial pneumonia, likely of viralaetiology. Moreover, a standing position is often main-tained in calves with breathing difficulties to facilitatebreathing. However, in a substantial proportion of ani-mals with high BALf neutrophil percentage these clinicalsigns were not demonstrated. A similar problem hasbeen reported in horses with inflammatory airway dis-ease, where BALf cytology is the only diagnostic means

Table 2 Final multivariable linear model for the association of clinical signs and broncho-alveolar lavage fluid characteristics withneutrophil percentage in broncho-alveolar lavage fluid of 335 group-housed calves

Variable Category Regression coefficient β (SE) P-value LSM estimate (SE)

Intercept 20.4 (0.14) < 0.001

P. multocida culture Negative Referent 22.7 (0.11)

Positive 0.32 (0.10) 0.005 30.3 (0.13)

Tracheal cough reflex Negative Referent 22.7 (0.10)

Positive 0.32 (0.15) 0.04 30.3 (0.16)

Posture Sternal recumbency Referent 21.4 (0.19)

Standing 0.47 (0.19) 0.03 32.1 (0.08)

Breathing frequency (breaths/min) 0.01 (0.0) 0.02 -

Random herd effect was significant (P < 0.001)

Table 1 Final multivariable linear model for the association of clinical signs and broncho-alveolar lavage fluid characteristics withtotal nucleated cell count (cells × 109 L) in broncho-alveolar lavage fluid of 332 group-housed calves

Variable Category Regression coefficient β (SE) P-value LSM estimate (SE)

Intercept 1.6 (0.14) < 0.001

Tracheal cough reflex Negative Referent 1.7 (0.04)

Positive 0.24 (0.07) 0.003 2.3 (0.08)

BALf blood staining Absent Referent 1.7 (0.04)

Present 0.26 (0.08) 0.002 2.4 (0.08)

Volume lavage fluid recovered (/mL) 0.03 (0.01) < 0.001 -

BALf broncho-alveolar lavage fluid. Random herd effect was not significant (P = 0.12)

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 8 of 12

when no abnormal clinical signs are evident [16]. Offcourse, continuous monitoring of clinical signs mighthave improved the diagnostic accuracy of these signs todetect lung inflammation in our study. A limitation ofthis study was that selection bias (convenience sampling)and potentially also classification bias cannot be ex-cluded, as all recordings, which are of subjective nature,were done by the same operator without blinding.The absence of a reference framework for nBAL cy-

tology was a serious limitation for our study. Given theabsence of good criteria defining a calf with healthy air-ways, this issue applies to all sampling techniques andthe cross-sectional design of this study appeared the bestoption to gain first insights. A clear finding was thatBALf neutrophil levels exceeded values previously re-ported in calves [21, 36, 37]. BALf neutrophil percentagehas been used as the primary indicator of inflammationand infection of the lower respiratory tract across spe-cies [7, 9, 37]. Suggested cut-off values defining BALfneutrophilia in calves are highly variable ranging from≥4 to 39% [21, 38]. Neutrophil percentage might be in-fluenced by the sampling volume and aliquot analysed.In human medicine, the first aliquot is often discardedsince it might contain a higher bronchial componentcompared to subsequent aliquots [17, 39]. As thecomplete BAL sample was used and the first aliquot wasnot discarded in our study, this might have played a role.

Furthermore, the bronchial component in BALf is in-creased when small volumes are used, as in this study,which could have resulted in an increased neutrophilpercentage [17, 20]. However, the nBAL catheter con-sisted of a diameter of 4 mm suggesting it would bewedged in a smaller bronchus deep in the lung, wherethe bronchial component is relatively reduced comparedto the volume of instilled fluid, diminishing the influenceof these technical factors. Nasal passage might also haveaffected BALf neutrophil percentage, especially in ani-mals with nasal discharge. We did not find any informa-tion in the literature regarding this aspect. In our studynasal discharge was not associated with increased BALfneutrophil percentage. Besides these technical aspects,both infection and inflammation due to non-infectiouscomponents of stable air, could explain the high neutro-phil percentage in this population. A limitation of thestudy was that for financial reasons extensive viral exam-inations were not included, which might have explaineda substantial proportion of the neutrophilia observed.In addition to the aforementioned factors, bacterial in-

fection would be a logical trigger of neutrophil influx,resulting in a very high neutrophil percentage [7–9]. Theuse of BAL in general, and nBAL specifically, for bacteri-ology has been controversial, mainly because of the po-tential risk of nasal contamination [24]. However, to theauthors knowledge, no studies providing evidence or

Table 3 Final multivariable linear model for the association of clinical signs and broncho-alveolar lavage fluid characteristics withmacrophage percentage in broncho-alveolar lavage fluid of 332 group-housed calves

Variable Category Regression coefficient β (SE) P-value LSM estimate (SE)

Intercept 57.7 (5.4) < 0.001

P. multocida culture Negative Referent 49.2 (2.1)

Positive - 4.5 (2.2) 0.04 44.7 (2.5)

Posture Sternal recumbency Referent 52.1 (3.7)

Standing - 10.2 (3.8) 0.007 41.8 (1.3)

Volume lavage fluidrecovered (/mL)

- 0.6 (0.27) 0.02 -

Random herd effect was significant (P < 0.001)

Table 4 Final multivariable linear model for the association of clinical signs and broncho-alveolar lavage fluid characteristics withlymphocyte percentage in broncho-alveolar lavage fluid of 335 group-housed calves

Variable Category Regression coefficient β (SE) P-value LSM estimate (SE)

Intercept 6.2 (0.34) < 0.001

Mentation Alert Referent 4.1 (0.10)

Depressed - 0.41 (0.19) 0.05 2.6 (0.20)

Age > 8 weeks Referent 4.2 (0.12)

4–8 weeks - 0.29 (0.11) 0.02 3.1 (0.03)

4 weeks - 0.41 (0.27) 0.15 2.7 (0.27)

BALf bloodstaining Absent Referent 2.8 (0.12)

Present 0.32 (0.12) 0.01 3.9 (0.16)

BALf broncho-alveolar lavage fluid. Random herd effect was significant (P < 0.001)

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 9 of 12

quantifying this risk of contamination are currentlyavailable. Mini-BAL procedures, comparable to thenBAL used in this study, have been used in humanmedicine and the interest in their potential use formicrobiology purposes has been increased [40–42] Incattle the nBAL has already been used frequently formicrobiology by practitioners, and gaining insights in itsusefulness for this purpose, appears urgent. In our data,a significant association between P. multocida and an in-creased neutrophil and decreased macrophage percent-age was demonstrated, as seen in another study [8].Taking the limitations of nBAL towards nasal passagecontamination into account, this might indicate that P.multocida could be a more virulent pathogen or a bettersecondary invader of airways that are already inflameddue to air pollution or viral infections, under the condi-tions of our study. In contrast to previous experimentalstudies, isolation of M. haemolytica was not linked toBALf neutrophil percentage, potentially due to differencesin virulence between strains [43, 44]. H. somni isolationwas not associated with neutrophilia, but increased bothlymphocyte and basophil percentage. This might point to-wards a different type of immune response against H.somni. Given the low number of H. somni and M. bovispositive cases, the absence of an association between isola-tion of H. somni and M. bovis and an increasing neutro-phil percentage should be interpreted with care.

Similar to BALf neutrophil percentage, TNCC countsexceeded values reported in previous work with the endo-scopic technique [21]. Cut-off values for TNCC have alsonot been established for calves, and they depend on thetechnique used [24, 45]. Our results showed that unlike thedifferential counts, TNCC values were mainly influenced bytechnical factors. TNCC increased with increasing amountof recovered lavage volume and macroscopically visibleblood staining. This blood staining was most likely causedby excessive vacuum (manual aspiration) or by the fragilityof inflamed respiratory mucosa.Eosinophil and basophil percentages in BALf from

young calves were studied for the first time in thispaper. Abnormal lung auscultation was associatedwith a BALf eosinophil percentage > 1%. Wheezes, asinduced by bronchoconstriction, were heard in 25%(6/24) of these calves. Next to parasitic infections,these findings could point towards the existence ofasthmatic syndromes in calves, as well-known inhumans and horses [16, 46, 47]. Moreover, basophilshave played a role in asthmatic syndromes and recentwork from human medicine showed that basophilswere activated in presence of suboptimal doses of al-lergens and bacteria [48]. Possibly, the observed asso-ciation between H. somni isolation and presence ofbasophils in BALf points towards a similar mechan-ism in calves.

Table 6 Final multivariable logistic regression model for the association of clinical signs and broncho-alveolar lavage fluidcharacteristics with eosinophil and basophil percentage in broncho-alveolar lavage fluid of 352 group-housed calves

Cell type Variable Category Calves(n =)

Percentage OR 95% CI P-value

Eosinophils a Lung auscultation Normal 159 3.8% Referent

Abnormal 193 8.8% 2.8 1–8.1 0.05

Basophils b H. somniculture

Negative 340 9.1% Referent

Positive 12 50.0% 1.4 2.9–45.5 < 0.001

BALf bloodstaining Absent 313 9.3% Referent

Present 39 20.5% 3.2 1.2–8.3 0.02

Lung consolidation > 3 cmin depth

Absent 245 7.3% Referent

Present 107 17.8% 2.6 1.2–5.6 0.01

BALf broncho-alveolar lavage fluid. Random herd effect was significant for eosinophils (P = 0.03) and not significant for basophils (P = 0.20)a Cut-off > 1% eosinophilsb Cut-off = ≥ 1 basophil counted on 400 cellsOR odds ratio, CI confidence interval

Table 5 Final multivariable linear model for the association of clinical signs and broncho-alveolar lavage fluid characteristics withepithelial cell percentage in broncho-alveolar lavage fluid of 335 group-housed calves

Variable Category Regression coefficient β (SE) P-value LSM estimate (SE)

Intercept 9.9 (0.25) < 0.001

Breathing frequency (breaths/min) - 0.01 (0.0) 0.03 -

Tracheal cough reflex Negative Referent 10.9 (0.07)

Positive - 0.53 (0.16) 0.004 6.3 (0.15)

Random herd effect was significant (P < 0.001)

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 10 of 12

Finally, an age effect for BALf cytological compositionhas been observed, characterized by higher lymphocyteand lower macrophage levels in BALf of older calves.The same age effects were demonstrated in peripheralblood [49]. The possibility, that the observed age effectfor lymphocytes was due to more frequent blood stain-ing of BALf in calves older than 8 weeks, cannot beexcluded in our study. In future studies this should beaccounted for when aiming to set cytology referencevalues for a certain BAL technique.

ConclusionsCytology findings as determined by an nBAL method,were only associated to a limited extent with ultrasono-graphic findings and selected clinical signs (positive tra-cheal reflex, standing position and breathing rate). BALfcytology offers additional information to the analysis ofrespiratory problems in calves, potentially aiding in bet-ter prevention and targeted treatment in the future. Inthis population, selected from herds not reporting anyrecent respiratory diseases, a high prevalence of lungconsolidation and animals with neutrophilia was de-tected, pointing towards issues with possible unaware-ness of the problem or subclinical disease presentations.

Additional files

Additional file 1: Table S1. Broncho-alveolar lavage fluid total nucleatedcell count and differential cell counts according to ultrasonographic lesionscore (ULS) based on 352 group-housed calves. (DOCX 26 kb)

Additional file 2: Table S2. Descriptives of broncho-alveolar fluid cellularcharacteristics from 352 group-housed calves. (DOCX 37 kb)

AbbreviationsBAL: Broncho-alveolar lavage; BALf: Broncho-alveolar lavage fluid;CI: Confidence interval; LSM: Least square means; nBAL: non-endoscopicbroncho-alveolar lavage; OR: Odds ratio; PPLO: Pleuropneumonia-likeorganism; R: Range; SD: Standard deviation; SE: Standard error; TNCC: Totalnucleated cell count; ULS: Ultrasonographic lesion score

AcknowledgementsWe would like to thank all the collaborating veterinarians and farmers fortheir help and interest in this project.

FundingThis work was done at the large animal clinic of Ghent University with aspecial research fund (BOF) of Ghent University, granted to K. van Leenen(01D25016). This research fund was used to cover costs of samplingmaterials and materials used for the analysis of the samples (bacteriologyand cytology). Bacterial species identification was done by MALDI-TOF MSfinanced by the Research Foundation Flanders (FWO-Vlaanderen) as Herculesproject AUGE/15/05 (G0H2516N).

Availability of data and materialsAll data generated or analysed during this study are included in thispublished article [and its Additional file 1 and Additional file 2].

Authors’ contributionsKvL generated the hypothesis and designed the experiment, organized andconducted the experiments, interpreted and analysed the results and wrotethe manuscript. BP generated the hypothesis, performed statistical analysis of

the data en interpreted the results, revised the manuscript and approved thefinal manuscript. LVD performed bacterial identification of bothPasteurellaceae and Mycoplasmata and assisted conducting the experiment.LDC assisted to collect the data and conducting the experiment. LG assistedto collect the data and conducting the experiment. CM assisted to collectthe data and conducting the experiment. FB revised the manuscript andinterpreted the results. PD revised the manuscript. All named authors haveread and approved the manuscript.

Ethics approval and consent to participateAll sampling techniques and the study protocol were revised and ethicalapproval to conduct the studies was obtained from the Ethical Committeeof the Faculty of Veterinary Medicine, Ghent University (ethical committeeapproval number EC2016–89). Calves were owned by individual farmers andinformed consent was obtained verbally from all owners of the animals.After the study all calves remained on the farm, no animals were sacrificedfor this study.

Consent for publicationNot applicable.

Competing interestsThe authors declare that they have no competing interests.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Author details1Department of Large Animal Internal Medicine, Faculty of VeterinaryMedicine, Ghent University, Salisburylaan 133, 9820 Merelbeke, Belgium.2Department of Pathology, Bacteriology and Avian Diseases, Faculty ofVeterinary Medicine, Ghent University, Salisburylaan 133, 9820 Merelbeke,Belgium.

Received: 11 March 2019 Accepted: 16 May 2019

References1. Adams EA, Buczinski S. Short communication: Ultrasonographic assessment

of lung consolidation postweaning and survival to the first lactation in dairyheifers. J Dairy Sci. 2016;99:1465–70.

2. Griffin D. Economic impact associated with respiratory disease in beef cattle.Vet Clin North Am Food Anim Pract. 1997;13:367–77.

3. Pardon B, Hostens M, Duchateau L, Dewulf J, De Bleecker K, Deprez P.Impact of respiratory disease, diarrhea, otitis and arthritis on mortality andcarcass traits in white veal calves. BMC Vet Res. 2013;9:79.

4. Pardon B, Catry B, Dewulf J, Persoons D, Hostens M, De bleecker K, et al.Prospective study on quantitative and qualitative antimicrobial and anti-inflammatory drug use in white veal calves. J Antimicrob Chemother. 2012;67:1027–38.

5. Cameron A, McAllister TA. Antimicrobial usage and resistance in beefproduction. J Anim Sci Biotechnol. 2016;7:68.

6. Threlfall J, Bergendahl H, Cocconcelli PS, Catry B, Girones R, Hederová J, etal. EMA and EFSA joint scientific opinion on measures to reduce the needto use antimicrobial agents in animal husbandry in the European Union,and the resulting impacts on food safety (RONAFA). EFSA J. 2017;15.

7. Johnson LR, Queen EV, Vernau W, Sykes JE, Byrne BA. Microbiologicand cytologic assessment of bronchoalveolar lavage fluid from dogswith lower respiratory tract infection: 105 cases (2001-2011). J VetIntern Med. 2013;27:259–67.

8. Allen JW, Viel L, Bateman KG, Rosendal S, Shewen PE. Cytological findings inbronchoalveolar lavage fluid from feedlot calves: associations withpulmonary microbial flora. Can J Vet Res. 1992;56:122–6.

9. Rossier Y, Sweeney CR, Rossier ZEL, et al. Bronchoalveolar lavage fluidcytologic findings in horses with pneumonia of pleuropneumonia. J Am VetMed Assoc. 1991;198:1001–4.

10. Holcombe SJ, Jackson C, Gerber V, Jercoat A, Berney C, Eberhardt S, et al.Stabling is associated with airway inflammation in young Arabian horses.Equine Vet J. 2010;33:244–9.

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 11 of 12

11. Robinson NE, Karmaus W, Holcombe SJ, Carr EA, Derksen FJ. Airwayinflammation in Michigan pleasure horses: prevalence and risk factors.Equine Vet J. 2006;38:293–9.

12. Pirie RS, Dixon PM, Collie DDS, McGorum BC. Pulmonary and systemiceffects of inhaled endotoxin in control and heaves horses. Equine Vet J.2010;33:311–8.

13. Sethi S. Infection as a comorbidity of COPD. Eur Respir J. 2010;35:1209–15.14. Hawkins EC, DeNicola DB. Bronchoalveolar lavage in the evaluation of

pulmonary disease in the dog and cat: state of the art. J Vet Intern Med.1990;4:267–74.

15. Hoffman AM. Bronchoalveolar lavage technique and cytological diagnosisof small airway inflammatory disease. Equine Vet Educ. 1999;11:330–6.

16. Couëtil LL, Cardwell JM, Gerber V, Lavoie JP, Léguillette R, Richard EA.Inflammatory airway disease of horses-revised consensus statement. J VetIntern Med. 2016;30:503–15.

17. Meyer KC, Raghu G, Baughman RP, Brown KK, Costabel U, du Bois RM, DrentM, Haslam PL, Kim DS, Nagai S, Rottoli P, Saltini C, Selman M, CharlieStrange BW. An official American Thoracic Society clinical practice guideline:the clinical utility of bronchoalveolar lavage cellular analysis in interstitiallung disease. Am J Respir Crit Care Med. 2012;185:1004–14.

18. Hoffman AM. Bronchoalveolar lavage: sampling technique and guidelinesfor Cytologic preparation and interpretation. Vet Clin North Am - EquinePract. 2008;24:423–35.

19. Hooi KS, Defarges AM, Sanchez AL, Nykamp SG, Weese JS, Abrams-OggACG, et al. Comparison of bronchoscopic and nonbronchoscopicbronchoalveolar lavage in healthy cats. Am J Vet Res. 2018;79:1209–16.

20. Sweeney C, Rossier Y, Ziemer E, Sweeney LS, et al. Effects of lung site andfluid volume on results of bronchoalveolar lavage fluid analysis in horses.Am J Vet Res. 1992;53:1376–9.

21. Ollivett TL, Caswell JL, Nydam DV, Duffield T, Leslie KE, Hewson J, et al. Thoracicultrasonography and Bronchoalveolar lavage fluid analysis in Holstein calveswith subclinical lung lesions. J Vet Intern Med. 2015;29:1728–34.

22. Van Driessche L, Valgaeren BR, Gille L, Boyen F, Ducatelle R, Haesebrouck F,et al. A deep nasopharyngeal swab versus nonendoscopic Bronchoalveolarlavage for isolation of bacterial pathogens from Preweaned calves withrespiratory disease. J Vet Intern Med. 2017;31:946–53.

23. Van Driessche L, Valgaeren B, De Schutter P, Gille L, Boyen F, Ducatelle R, etal. Effect of sedation on the intrapulmonary position of a bronchoalveolarlavage catheter in calves. Vet Rec. 2016;179:2016–9.

24. Hodgson JL. Collection and interpretation of tracheal wash andbronchoalveolar lavage for diagnosis of infectious and non-infectious lowerairway disorders. In: Proceedings of the 9th International Congress of WEVA.Marrakech, Morocco; 2006. p. 71–77.

25. Kollef MH, Bock KR, Richards RD, Hearns ML. The safety and diagnosticaccuracy of Minibronchoalveolar lavage in patients with suspectedventilator-associated pneumonia. Ann Intern Med. 1995;122:743–8.

26. Tasbakan MS, Basoglu OK, Ekren K. Comparison of Bronchoalveolar lavageand mini-Bronchoalveolar lavage in the diagnosis of pneumonia inimmunocompromised patients. Respiration. 2011;81:229–35.

27. Lavigne MC. Nonbronchoscopic methods [NonbronchoscopicBronchoalveolar lavage (BAL), mini-BAL, blinded bronchial sampling, blindedprotected specimen brush] to investigate for pulmonary infections,inflammation, and cellular an molecular markers: a narrative review. ClinPulm Med. 2017;24:13–25.

28. Ollivett TL, Buczinski S. On-farm use of ultrasonography for bovinerespiratory disease. Vet Clin North Am Food Anim Pract. 2016;32:19–35.

29. Babkine M, Blond L. Ultrasonography of the bovine respiratory system and itspractical application. Vet Clin North Am Food Anim Pract. 2009;25:633–49.

30. Buczinski S, Forté G, Bélanger A. Short communication: Ultrasonographicassessment of the thorax as a fast technique to assess pulmonary lesions indairy calves with bovine respiratory disease. J Dairy Sci. 2013;96:4523–8.

31. Shimizu T. Selective medium for the isolation of mycoplasma bovis fromnasal discharges of pneumonic calves. Res Vet Sci. 1983;34:371–3.

32. Meyer KC. Bronchoalveolar lavage as a diagnostic tool. Semin Respir CritCare Med. 2007;28:546–60.

33. Olsen HH, Grunewald J, Tornling G, Sköld CM, Eklund A. Bronchoalveolarlavage results are independent of season, age, gender and collection site.PLoS One. 2012;7(8).

34. Teixeira AGV, McArt JAA, Bicalho RC. Thoracic ultrasound assessment oflung consolidation at weaning in Holstein dairy heifers: reproductiveperformance and survival. J Dairy Sci. 2017;100:2985–91.

35. Lekeux P, Amory H, Desmecht D, Gustin P, Linden A, Rollin F. Oxygen transportchain in double-muscled blue belgian cattle. Br Vet J. 1994;150:463–71.

36. Pringle JK, Viel L, Shewen PE, Willoughby RA, Martin SW, Valli VE.Bronchoalveolar lavage of cranial and caudal lung regions in selectednormal calves: cellular, microbiological, immunoglobulin, serological andhistological variables. Can J Vet Res. 1988;52:239.

37. Allen JW, Viel L, Bateman KG, Rosendal S. Changes in the bacterial flora ofthe upper and lower respiratory tracts and bronchoalveolar lavagedifferential cell counts in feedlot calves treated for respiratory diseases. CanJ Vet Res. 1992;56:177–83.

38. McGuirk SM. Disease Management of Dairy Calves and Heifers. Vet ClinNorth Am Food Anim Pract. 2008;24:139–53.

39. Pohunek P, Pokorná H, Stříž I. Comparison of cell profiles in separatelyevaluated fractions of bronchoalveolar lavage (BAL) fluid in children. Thorax.1996;51:615–8.

40. Pugin J, Auckenthaler R, Mili N, Janssens J-P, Lew PD, Suter PM. Diagnosis ofventilator-associated pneumonia by bacteriologic analysis of Bronchoscopicand Nonbronchoscopic “blind” Bronchoalveolar lavage fluid. Am Rev RespirDis. 1991;143:1121–9.

41. Humphreys H, Winter R, Baker M, Smith C. Comparison of bronchoalveolarlavage and catheter lavage to confirm ventilator-associated lowerrespiratory tract infection. J Med Microbiol. 1996;45:226–31.

42. Flanagan PG, Findlay GP, Magee JT, Ionescu AA, Barnes RA. The diagnosis ofventilator-associated pneumonia using non-bronchoscopic , non-directedlung lavages. Intensive Care Med. 2000;26:20–30.

43. Weiss OJ, Bauer MC, Whiteley LO, Maheswaran SKAT. Changes in blood andbronchoalveolar lavage fluid components in calves with experimentallyinduced pneumonic pasteurellosis. Am J Vet Res. 1991;52:337–44.

44. Lopez A, Maxie MG, Ruhnke L, Savan M, Thomson RG. Cellular inflammatoryresponse in the lungs of calves exposed to bovine viral diarrhea virus,mycoplasma bovis, and Pasteurella haemolytica. Am J Vet Res. 1986;47:1283–6.

45. Lay JC, Slauson DO, Castleman WL. Volume-controlled bronchopulmonarylavage of Normal and pneumonic calves. Vet Pathol. 1986;23:673–80.

46. De Giacomi F, Decker PA, Vassallo R, Ryu JH. Acute eosinophilicpneumonia: correlation of clinical characteristics with underlying cause.Chest. 2017;152:379–85.

47. Ivester KM, Couëtil LL, Moore GE, Zimmerman NJ, Raskin RE. Environmentalexposures and airway inflammation in young thoroughbred horses. J VetIntern Med. 2014;28:918–24.

48. Melioli G, Passalacqua G, Baena-Cagnani CE, Canonica GW. Allergens andbacteria interaction in the induction of basophil activation: is this the lostring between allergy and infections in pediatric patients? Curr Opin AllergyClin Immunol. 2012;12:164–70.

49. Brun-Hansen HC, Kampen AH, Lund A. Hematologic values in calves duringthe first 6 months of life. Vet Clin Pathol. 2006;35:182–7.

van Leenen et al. BMC Veterinary Research (2019) 15:167 Page 12 of 12